Siadhal Magos and Shahriar Tajbakhsh were both working in the tech industry, at Uber and Palantir respectively, when they noticed a common struggle for many corporate HR departments: the hiring process. Particularly, the process of interviewing was becoming overwhelming and complicated.
Magos told TechCrunch, “It was clear to us that interviews were the most important aspect of the hiring process, but also the most unreliable and opaque. Not to mention, there was a lot of tedious note-taking and feedback writing that many interviewers and hiring managers tried to avoid.”
Magos and Tajbakhsh saw an opportunity to disrupt the hiring process, but they wanted to maintain the human element. So, they created Metaview – an AI-powered note-taking app for recruiters and hiring managers to record, analyze, and summarize job interviews.
“Metaview is specifically designed for the hiring process,” Magos explained. “It allows recruiters and hiring managers to focus on getting to know candidates rather than extracting data from conversations. As a result, it saves time and allows interviewers to be fully present during interviews.”
The platform integrates with various apps, phone systems, and video conferencing platforms, as well as tools like Calendly and GoodTime. This allows for automatic capturing of interview content. The platform also takes into account the nuances of recruiting conversations and uses data from other sources, such as applicant tracking systems, to highlight the most relevant moments.
According to Magos, platforms like Zoom, Microsoft Teams, and Google Meet all have built-in transcription features, but they don’t cater specifically to the recruiting use case. Metaview’s AI is able to extract more relevant information from interviews and also assists users with their recruiting workflows.
Although traditional job interviews have their flaws, tools like Metaview could potentially help improve the process. As noted in Psychology Today, our brains are often biased and can hinder our judgement and decision-making. For example, we tend to rely too heavily on the first piece of information offered and interpret data in a way that confirms our preexisting beliefs.
However, the question remains: does Metaview truly work for all users?
Unfortunately, biases still exist even in the best AI-powered systems. A study by Stanford University showed that speech-to-text services from major companies have significantly higher error rates for Black speakers compared to white speakers. Additionally, a more recent study published in the journal Computer Speech and Language found significant differences in how two leading speech recognition models treat speakers of different genders, ages, and accents.
There is also the issue of AI-generated “hallucinations.” In a recent incident reported by The Wall Street Journal, Microsoft’s AI Copilot tool created false attendees and subjects that were never discussed in a meeting summary.
When asked about Metaview’s efforts to mitigate bias and other algorithmic issues, Magos stated that the platform’s training data is diverse enough to exceed human performance in recruitment workflows and has achieved success on popular benchmarks for bias.
However, there may be cause for concern in how Metaview handles speech data. According to Magos, the platform stores conversation data for two years by default, unless users request for it to be deleted. This may seem like a long time, and potential job candidates might have reservations.
Despite these concerns, Metaview continues to gain both funding and clients. In fact, the London-based startup recently raised $7 million from investors including Plural, Coelius Capital, and Vertex Ventures, bringing their total raised amount to $14 million. Magos reveals that Metaview has 500 clients, including well-known companies such as Brex, Quora, Pleo, and Improbable, and has seen a 2,000% growth year-over-year.
“The funds will primarily be used to expand our product and engineering teams and fuel our sales and marketing efforts,” says Magos. “We plan to triple our product and engineering team in order to further enhance our conversation synthesis engine, ensuring that our AI is able to extract exactly the information our clients need. We also aim to develop systems that can proactively detect issues, such as inconsistencies in the interview process or candidates losing interest.”
